LAUSR

Particle imaging has shown growing interest in the recent literature as it allows direct measurement of the tissue relative stopping power (RSP) within a patient, using either the radiographic or tomographic images [1] . However, its accuracy is limited by the multiple Coulomb scattering (MCS) suffered by the particles through the phantom and by the secondary particles noise measured on the detector that creates artefact in the reconstructed images. Monte Carlo (MC) techniques help investigate the spatial resolution gain obtained by using 1) different path estimate or 2) different ions to reduce the MCS effect on the final images [2] , [3] , [4] . Furthermore, secondary particles filtering techniques are paramount to achieve high RSP accuracy but may reduce contrast and spatial resolution along the process. The effect of the applied filters can be directly studied on the noise power spectrum (NPS) of perfect images acquired through MC simulations. Initial investigation demonstrates that He ions is the optimal choice between light ions (i.e. 1H to 6C) to achieve the highest path accuracy [2] decreasing the root-mean square uncertainty between the MC reference and the path estimate from 0.5 [mm] for protons to 0.18 [mm] for He. This has been shown to leads to higher spatial resolution radiographic images. On the other hands, the NPS spectrum [mm2] increases due to secondary particles on unfiltered He image is 300 times higher than that of proton, stressing the need for smart secondary particle filters.